Re: New module proposal: LightDM

[Robert Ancell <robert ancell gmail com>]

On 13 May 2011 18:32, Ray Strode <halfline gmail com> wrote:
> Hi,
>
> (speaking again as one of the three GDM maintainers)
>
> On Fri, May 13, 2011 at 8:59 AM, Robert Ancell <robert ancell gmail com> wrote:
>> Why replace GDM?
>>
>> - LightDM is a cross-platform solution.
> What platforms does LightDM support that GDM doesn't?  Are they
> platforms GNOME is targetting?  Not sure this is necessarily a win
> without knowing more details.

The value to GNOME is not that it can run on different platforms, the
value is that the component can be shared.

>> Ubuntu is planning to switch to it this cycle,
> Sure, and I can see how that decision may potentially make sense for
> Canonical (you would be able to offer in-house expertise, better
> control over unity integration etc etc.)

And we would have loved to use GDM, it certainly wasn't through lack
of trying.  For Ubuntu the Unity aspects will be in the Unity greeter,
but the core LightDM daemon/libraries will not contain anything Unity
specific.

>>  By sharing this piece of infrastructure GNOME can spend
>> more time working on important GNOME components.
> So you're saying, we could stop working on GDM, Brian, McCann and I
> could leave the login screen to you to handle and we could use the
> extra time we got to work on other parts of GNOME?  There's no
> question that we're all busy, for sure, but let's look at the "big
> things" (modulo small patch review, security errata, etc, that would
> exist in both projects) needed for GDM:
>
> 1) Landing the multiple simultaneous pam conversations branch of GDM
> so you get sane behavior in the presence of smartcards, fingerprint
> readers, etc.  This lets you swipe your finger, insert your smart
> card, or whatever while sitting at a enter password prompt.
> 2) Giving GDM a more of a GNOME 3 look and feel (as per the mockups
> you already mentioned elsewhere in the thread)
> 3) Better support starting a login screen dynamically a keyboard,
> mouse, and display show up ("multi-seat").  Also, see the work Lennart
> and Kay are discussing lower in the stack to help facilitate some of
> this.
> 4) Figuring out how GDM,  screen locking, and the shell all fit together
>
> I don't think LightDM will help us with any of these 4, will it?

These are all valid features and I completely support making a list of
requirements for GNOME for LightDM to be even considered.  I am
absolutely fine for this proposal to be rejected on grounds like this.

>> Note that LightDM is not lighter in features, but in architecture.
> And a different focus, right? GDM is firmly a GNOME project, designed
> to integrate and work well with GNOME.  LightDM is designed with the
> idea of being more generic right?

More generic in the parts that are common.  In the parts that are
GNOME specific, as differentiated as is required.

>> - By having a well defined interface between the greeter and daemon,
>> it is significantly easier to develop a greeter without knowledge of
>> how display management works.  This is useful as the skillset and
>> motivations of these two sets of developers are different.
> Not sure how much of a selling point "multiple greeters" is, but GDM's
> architecture allows for it.
>
> Dr. Mo even did one apparently:
> http://doctormo.org/2011/04/12/how-to-make-a-gnome-login-screen-in-python/

Yes, I did congratulate him on that! :)

> Anyway, I'm obviously of the opinion we should stick with GDM.  There
> just doesn't seem to be a good reason to switch.

I'll keep trying to convince you, thanks for the feedback Ray. :)